A nitrogen budget for Denmark; developments between 1990 and 2010, and prospects for the future

A nitrogen (N) budget for Denmark has been developed for the years 1990 to 2010, describing the inputs and outputs at the national scale and the internal flows between relevant sectors of the economy. Satisfactorily closing the N budgets for some sectors of the economy was not possible, due to missing or contradictory information. The budgets were nevertheless considered sufficiently reliable to quantify the major flows. Agriculture was responsible for the majority of inputs, though fisheries and energy generation also made significant contributions. Agriculture was the main source of N input to the aquatic environment, whereas agriculture, energy generation and transport all contributed to emissions of reactive N gases to the atmosphere. Significant reductions in inputs of reactive N have been achieved during the 20 years, mainly by restricting the use of N for crop production and improving livestock feeding. This reduction has helped reduce nitrate leaching by about half. Measures to limit ammonia emissions from agriculture and mono-nitrogen oxides (NOx) emissions from energy generation and transport, has reduced gaseous emissions of reactive N. Much N flows through the food and feed processing industries and there is a cascade of N through the consumer to solid and liquid waste management systems. The budget was used to frame a discussion of the potential for further reductions in losses of reactive N to the environment. These will include increasing the recycling of N between economic sectors, increasing the need for the assessment of knock-on effects of interventions within the context of the national N cycle.


Introduction
The benefits of reactive nitrogen (N) for food production and industrial products, and the threat posed to human and ecosystem health by losses of reactive nitrogen were well described in the European Nitrogen Assessment . At the European scale, an economic assessment has suggested that the balance between cost and benefits is Environmental Research Letters Environ. Res. Lett. 9 (2014) 115012 (8pp) doi: 10.1088/1748-9326/9/11/115012 Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. not optimal and that further action is necessary to redress the balance (van Grinsven et al 2013). Over the past 25 years, Denmark has introduced increasingly stringent environmental legislation to curb emissions of ammonia (NH 3 ) and NO x , and losses of nitrate ( − NO 3 ) by leaching. The national legislation often exceeds (and in some cases predates) international agreements and legislation such as the UNECE Convention on Long-Range Transboundary Air Pollution (CLRTP) (UNECE 1979) and the European Union Nitrates Directive (EU 1991). The major sources of the N pollution have been agriculture (NH 3 and − NO 3 ), energy production and transport (NO x ). Measures with the least cost have been implemented and future progress will be achieved at an increasing marginal cost. It is therefore worthwhile investigating other economic sectors to identify opportunities for increasing the efficiency of use of reactive N and reducing N losses.
The construction of partial N budgets for particular sectors such as agriculture (Parris 1998) or at the national scale (Slak et al 1998) is considered a useful way to identify major imbalances between the input and output of N at a range of scales. Fleckseder (1992) proposed using national N budgets as an aid to policymaking but the idea has lain dormant until recently, when Leip et al (2011) suggested that such budgets would be a valuable tool in resource optimization, not least as a policy tool for identifying opportunities for intervention. There are indications that this suggestion is being adopted (Junker 2013). Wier and Hasler (1999) developed a national terrestrial/aquatic/atmospheric N budget for Denmark for a single period in the late 1980s. Here, we report a first attempt at constructing a national N budget for Denmark for the years 1990 to 2010 and then use it as a framework for discussing opportunities for achieving a further reduction in losses of reactive N.

Methods
The sources of data used for the construction of the Danish N budget are documented in detail in the supplementary material, available at stacks.iop.org/ERL/9/115012/mmedia. Data on the mass flows of commodities are primarily collated from the statistical data held by Statistics Denmark. Data concerning losses of N species to the environment were collated from reports prepared to document Denmark's compliance with the United Nations Framework Convention on Climate Change and the United Nations Convention on Long-Range Transboundary Air Pollution UNECE (1979), supplemented by reports prepared to support national policy-making. The following major pools involved in the N budget of Denmark were identified: • Agriculture, with subdivision into livestock and crop production. • Fisheries.
• Forestry and semi-natural areas.
• Food and feed processing industries.
• Consumers and urban areas.
The N concentrations in commodities were obtained from a wide range of sources. Where possible, data was collated for all years; where not, effort was made to collate data at five yearly intervals. Where data were missing, values were extrapolated from other years. In a number of cases, it was not possible to satisfactorily close the budget, either because data were not available or because there was a conflict between different data sources. Where these cases involved significant N flows, they are identified and discussed below. The data sources are detailed in the supplementary material.

Methodological issues
Assessing the flow of N in feed to ruminant livestock proved problematic. The statistics concerning livestock feed are the sum of imported and domestically-produced feedstuffs. The statistics concerning feed imports can be considered reliable. The statistics covering domestic crop production are for grain and other crop products sold are based on documented material flows, whereas for roughage crops, production is based on estimates made by agricultural advisors. The latter accounts for 40-50% of the total N harvested in crops and a larger proportion of the animal feed, so the statistics concerning this major N flow are largely based on expert judgment. Using the official statistics for N in crop products used for animal feed and subtracting the N exported in animal products or estimated to be lost by gaseous emission from the manure management system, yields an estimate of N applied to land in animal manure. These estimates proved to be on average 24% higher than the official values used when estimating field emissions of NH 3 and N 2 O, and the leaching of − NO 3 . The official manure application values are calculated from a survey of commercial feed rations, the statistics on the production of animal products, and estimates of N lost by gaseous emission from the manure management system. The method used to calculate the official manure application values was considered the more robust, so the consumption of N in animal feed was estimated here as the sum of the N applied in manure and the gaseous N losses. The official crop production statistics were nevertheless used for calculating N flows associated with crop and soil, since no alternative was available. Since the change in soil N was calculated as the difference between the inputs and outputs/losses, this means that errors in the estimate of crop production will propagate to this item. Note that for the N input in crop products to Food/ feed processing, the estimates of Vinther and Olsen (2013) were used, since they identified the net export of crop products from agriculture.
Estimating the N produced as roughage feed for ruminant livestock is a recurrent problem when constructing N budgets for ruminant livestock farming, e.g. when calculating gross N balances using the method recommended by OECD (Parris 1998). Representative field surveying would be difficult and expensive and expert judgement represents a poor substitute for objective data, while basing roughage crop production on a survey of commercial feeding practices would be more reliable. Livestock farming will be a key source of greenhouse gas emissions for most countries, so they should be using the IPCC Tier 2 methodology (IPCC 2006) or a Tier 3 methodology for estimating N intake for ruminant livestock. This requires an estimate of the energy and N content of livestock rations, so representative feed rations should be available.
Estimating the N flow in food/feed processing and waste streams proved difficult. It was not possible to satisfactorily close the N budget for food/feed processing, with the difference between the N input and output varying between −9% and 20% of the amount input. The difference between the amount estimated to be input to the sewage system by consumers and the amount estimated to enter waste water treatment plants averaged 25% of the former. In contrast, for waste water treatment plants and solid refuse treatment, the amount estimated to be output exceeded the amount input by averages of 27% and 25% respectively. The N flows within and between the agricultural and aquatic ecosystems has been a focus of much research in Denmark over the last 20 years, due to the political attention given to − NO 3 leaching. In contrast, nutrient flows within sectors dealing with food waste and waste management have only recently received political attention and this is reflected in the quality of the statistical information available; the quality and detail of the data collected and collated at the national scale depends primarily on the demand for information from policymakers.
Since the questions of interest to policymakers change with time, together with the resources available and methodologies used, it will always be difficult to obtain complete and consistent data for a 20 year period. Furthermore, data are usually collected on a sector by sector basis, sometimes creating difficulties quantifying inter-sectorial material flows. If a national N budget for Denmark is to be of value for policy-making or informing the public debate, further work is required to resolve some of the difficulties found in this work. Since the exercise undertaken here exclusively accessed information publicly available via the Internet, some of which was sparsely documented, it is likely that many of the issues can be resolved by reference to information only available on paper or by interacting with the data providers.

National budget
The national N budgets for Denmark in 1990 and 2010 are shown in figure 1. Note that to reduce clutter, a number of minor flows have been omitted. Detailed figures and data, including flows for intermediate years, are presented in the supplementary material. The main changes between 1990 and 2010 are as follows. The import of N in air from foreign sources decreased, due to the reductions in emissions by other countries, under CLRTP. The export of N in air from Denmark has also reduced, due to both CLRTP and national legislation. The input of N to the energy sector increased but the emission of NO x decreased, again due to emission controls. The emission of NO x from the transport sector also reduced, due to the introduction of catalytic converters.
In both years, much of the N cycling in Denmark was associated with agriculture. The input of N into livestock production in in feed from domestic sources decreased slightly but the use of imported feed increased, so that the total input to livestock increased. The production of animal products increased, partly as there were more livestock but also due an increase in the efficiency of N the livestock; this effect that can also be seen in a slight decrease in the amount of N excreted (see supplementary material). The requirement for farmers to take account of the available N in manure when fertilizing crops meant that the import of N fertilizer reduced considerably (yellow arrow in figure 1). Together with the slight reduction in N added in animal manure, this meant the total application of N to the soil was reduced considerably. Emissions of N 2 O and N 2 from cropland decreased accordingly. Losses of NH 3 from agriculture were reduced during this period, due to regulations imposing the use of low emission technology in the manure management system, the reduction in N fertilizer applications and a ban on the treatment of straw with NH 3 to improve its feed value. A large reduction in − NO 3 leaching to groundwater was also achieved, partly through the reduction of N input to the soil but also through regulations concerning the minimum requirements for winter cropping. The reduction of − NO 3 leaching to groundwater results in reduced input to the remainder of the aquatic ecosystem. In addition to reducing the eutrophication of coastal waters, there is an additional reduction in the N 2 O emissions as less − NO 3 is available in the surface waters.
The input of N to Food processing/Distribution remained large, with most of the N being exported. There was an increase in the flow associated with animal products but a slight decrease in the flow in crop products (which here includes animal feed for export). N flows were smaller in other sectors and the magnitude of the changes in time less certain (see above). However, it is clear that there has been a reduction in the discharge of N in sewage effluent to coastal waters and an increase in emissions of N 2 from sewage treatment, due to better wastewater treatment.

Comparison with other studies
The values obtained in the current study are broadly similar to those for the late 1980s from Wier and Hasler (1999), although differences in the structure of the budgets make a direct comparison for some flows difficult. The deposition of N to terrestrial and aquatic ecosystems is evaluated somewhat higher, as are emissions of reactive and non-reactive gases from agriculture. Some differences would be expected, partly due to the different periods, partly because limitations in some statistical information leave scope for different interpretations.
The N flows into, within and out of a country would be expected to vary according to a number of factors; the area and quality of land influences the agricultural production and the human population influences the proportion of that production consumed domestically and the importance of nonagricultural activities. Globally, the level of economic development is likely to affect the proportion of the population living in urban areas and the consumption of animal products. Of the national N budgets currently available, the budget for The Netherlands published in Leip et al (2011) is the one with which it is most appropriate to compare the results of this study; a Western European country with an intensive agricultural industry. The data in Leip et al (2011) is for The Netherlands in the 1990s. In this period, the similarities with Denmark for the same period are the large flow of N in fertilizer and animal feed, and the losses of reactive N gases. The main differences are that for the case of The Netherlands, the manure production was higher, the use of industrial and agricultural products by consumers was higher, the leaching of − NO 3 from agriculture was lower and the loss of N 2 from agriculture much higher. The manure production appears to be particularly high, considering that the agricultural area was about one third of that of the Danish, reflecting the importance and intensity of the livestock sector in The Netherlands at that time. The difference in consumption of agricultural products can be explained by the difference in population; consumption was estimated to be about 7.9 and 6.7 kg N capita −1 yr −1 for The Netherlands and Denmark respectively. In contrast, the consumption of 20 kg N capita −1 yr −1 of non-food commodities in The Netherlands is difficult to explain and requires further investigation. The − NO 3 -N leaching in The Netherlands was equivalent to about 46 kg N (ha agricultural area) −1 yr −1 , which is about half that of Denmark at the time (about 93 kg N (ha agricultural area) −1 yr −1 ). Conversely, the loss of N 2 of 210 kg N (ha agricultural area) −1 yr −1 in The Netherlands was much higher than in Denmark (28 kg N (ha agricultural area) −1 yr −1 ). In much of The Netherlands, the water table is close to the surface of agricultural land, which leads to high emissions via denitrification and a consequent reduction in the proportion of − NO 3 created that remains for leaching. In contrast, the water table is much lower in most of Denmark, so denitrification is lower and a greater proportion of the − NO 3 created remains for leaching. Substantial losses of N 2 via denitrification of agricultural − NO 3 also occur in Denmark but are here attributed to the aquatic environment rather than agriculture.
3.4. Technical measures to reduce harmful N losses to the environment 3.4.1. Agriculture. Inputs in the form of fertilizer and animal feed to support agricultural production are responsible for much of the N that drives harmful losses, particularly − NO 3 leaching and emissions of NH 3 and N 2 O. However, driven by a combination of legislation and economic advantage, Danish farmers have already achieved substantial reductions in N losses, through improved animal and crop husbandry (Dalgaard et al 2014). This combination seems likely to maintain the reductions achieved, even in the face of moderate changes in the market (e.g. the removal of milk quotas). On the other hand, the least cost (and sometimes negative cost) measures are largely already implemented, so further improvements will be challenging.
A number of additional measures are currently being implemented. Restrictions on NH 3 emissions and the enforced N fertilization of crops to 15-20% below the economic optimum are together encouraging farmers to employ low NH 3 emission technology. The aim of the Danish government is to double the area farmed organically by 2020, from the current value of about 7% of the agricultural area. If successful, this may reduce N losses, since the N surplus for Danish organic farms appears to be lower than for conventional farming (Dalgaard et al 1998). As part of its future energy strategy (Danish Ministry of Climate, Energy and Building 2012), the Danish government has a target of anaerobically digesting 50% of animal manure by 2020 and although the primary objective of the policy is the reduction in fossil fuel usage, a side-effect is that the mineralization of manure organic N will improve the efficiency with which it can be used for fertilization. Under Danish nutrient legislation, this will result in a further reduction of mineral fertilizer use. Following the report of a commission on agriculture and nature (Nature and Agriculture Commission 2013) and in response to the EU Water Framework Directive (EU 2000), a second measure under development is to require land that is particularly susceptible to − NO 3 leaching to be used for crops that require little N fertilization or have the ability to retain a high proportion of N applied (e.g. woody perennials for energy production), or to be taken out of agricultural production completely. At present, it appears that the value of energy crops is too low and the price of land too high for this to be economically feasible for implementation over large areas. In some locations, it may be feasible to intercept leached − NO 3 before it can enter watercourses, either by creating artificial wetlands (although the focus of the latter is currently removal of P rather N) or by growing, harvesting and removing an under-fertilized perennial crop such as grass. The latter option is feasible if there is a local recipient for the grass, such as a ruminant livestock farm, and the quality of the grass is suitable. In the longer term, it is possible that the development of a biorefinery capacity will create an additional market for such material and provide an additional route for recycling N.
3.4.2. Transport and energy. The opportunities for further reducing the NO x emissions from transport and energy via abatement technologies are limited, since much has already been achieved. However, changing to electricity as the source of power would reduce NO x emissions, provided the electricity were generated using a method that did not itself lead to higher NO x emissions (e.g. hydroelectric, solar, wind).
3.4.3. Food/feed processing. Significant steps have already been taken to minimize the waste from food and feed processing. Carcass parts of slaughtered livestock not commonly consumed in western cultures are marketed in the Far East or converted into feed for pets or fur animals. Readily-degradable carcass fractions such as the gut contents are also used as feedstock for biogas generation or for biofuel production, while other fractions are used as organic fertilizer. The potential for further efficiencies appears limited here. However, there remains scope for reducing the food waste from retailers, from restaurants and in institutional food preparation (Mogensen et al 2013), and recycling that which remains.
3.4.4. Consumers, refuse and wastewater. Food waste by consumers remains a significant issue that can be tackled by education (Mogensen et al 2013). Food waste cannot be eliminated completely, since some foods are sold in forms that contain inedible components (e.g. citrus fruit peel, bones). Reducing food waste would not reduce N losses unless it resulted in reduced the amount of food purchased. If the effect were to be to increase consumption, the N would be partitioned to wastewater as excreta rather than solid domestic refuse. At the national scale, a reduction in food purchases would probably result in lower imports of food not produced in Denmark and higher exports of those that are.
Most of the N in refuse is contained in food or garden waste. At present, much of Danish refuse is used for energy generation, without separation. Significant progress can therefore be made towards recycling the N via composting or anaerobic digestion and the government has established targets for substantially increasing the recycling of organic waste by these means (Danish Ministry of the Environment 2013).
Most domestic sewage and much industrial sludges are already recycled as organic manures to agriculture. However, the choice of aerobic treatment followed by denitrification means that most of the N entering wastewater treatment plants is lost to the atmosphere as N 2 , rather than being recycled. In the future, it would be preferable if this N were to be captured, reprocessed and recycled (Oenema et al 2011).

Strategic measures
One of the major ecological impacts of reactive N is the eutrophication of seminatural ecosystems. In terrestrial seminatural ecosystems, deposition of N leads to unwanted changes in plant species composition whereas in coastal waters, it is a major contributor to algal blooms and subsequent anoxia. For Denmark, N deposition resulting from domestic reactive N emissions only accounts for only about 33% of deposition to land and 12% of deposition to coastal waters, the remainder coming from other countries (Jensen et al 2012). At the same time, Denmark has at significant cost implemented some of Europe's most stringent measures to limit NH 3 and NO x emissions. Denmark therefore has a strong ecological and economic interest in pressing neighboring countries to commit to ambitious N emission reductions and subsequently monitoring compliance with these commitments.
Two initiatives currently on the policy horizon could have a significant impact on N flows Danish crop production in the longer term. The development of a substantial biorefinery capacity (Jørgensen 2014), aimed at replacing industrial feedstocks currently provided from fossil materials, could lead to a significant extension in the area planted to perennial crops. Protein-rich byproducts from biorefineries could provide an alternative to imported protein-rich feeds such as soya. However, under existing nutrient management legislation, a switch from annual to perennial cropping would result in an increase in the amount of plant-available N per hectare that could be applied to the land cropped in this way. The second development that could impact Danish crop production would be the development of biomass production for energy generation (Jørgensen et al 2013). In contrast to crop production for biorefineries, energy crops tend to have a moderate to high ability to retain N; moderate for maize silage, high for perennials such as willow coppice and Miscanthus spp.
Livestock farming has been particularly associated with high losses of reactive N (Leip et al 2014), leading to a suggestion that measures to reduce the amount of animal protein in human diets would be beneficial (Reay et al 2011). Refocussing agriculture away from livestock production and towards crop production has also been suggested as a means of increasing the global availability of protein for human consumption (Aiking 2014). This may be true if livestock are only fed crop products not suitable for human consumption (Fairlie 2010). This may be true from a global perspective but from a purely Danish perspective, the environmental effect of such a change in domestic consumption patterns is likely to be limited since any reduction would probably result in increased exports.
The direct impact of any policy-driven change in Danish food production on global food supply will be limited. Denmark accounts for respectively about 0.2, 1.5 and 0.8% of global beef, pork and milk production (FAOSTATS 2014). The contribution to food exports is greater; about 3.5, 5.1, and 2% of cow butter, cheese and other dairy products, 2.7% of beef and 9.7% of pork (FAOSTATS 2014). In contrast, the direct impact of any policy-driven change in Danish food production on Denmark is potentially high; agriculture and food production accounts for 12% of Danish gross domestic product, employs 8.5% of the labor force and the export of agricultural and food products accounts for 12% of Danish exports (Danish Agriculture and Food Council 2014). Since the low-cost or negative-cost measures to reduce the circulation of reactive N have long since been implemented, it is useful to consider why the Danish government continues to take a positive approach to reducing this circulation. The first reason is domestic public pressure to limit the environmental damage caused by reactive N. As a consequence, the policy of the Ministry of Agriculture, Food and Fisheries is to balance the economic benefits and environmental costs of farming (Ministry of Agriculture, Food and Fisheries 2014). The second appears to be the belief that by imposing more stringent regulation, it will create a domestic market for environmental technologies, thereby encouraging research and development that will lead to the sale of equipment and expertise to other countries (Danish Energy Agency 2014). This strategy presupposes that there is a long-term global (or at least European) trend to restrict N emissions to the environment and that any short-term disadvantage to Danish agriculture can be offset by government subsidies. Whilst the latter is under the control of the Danish government, the former requires international acceptance of the need for further measures to reduce losses of reactive N. We believe that national N budgets provide an informative and persuasive method of convincing both public and politicians of the need to continue this momentum.

Value of a national approach
The prominence of agriculture in national N budgets is a common feature in the studies available (Leip et al 2011). The prominence of agriculture in Denmark and the magnitude of the N flows into, out of and within the sector has meant and will continue to mean that significant progress in reducing the losses of reactive N to the environment can be made by considering this sector alone. The value of taking a national, multi-sectoral approach, beyond that of the educational value, has to be argued in terms of achieving a more cost-effective reduction by taking a holistic approach and gaining an understanding of the flows between sectors. The benefits of a cross-sectoral approach are already proven in Denmark, by the work of agricultural, atmospheric and hydrological scientists to assess the contribution of different sources to the input of N to coastal waters.

Conclusions
Identifying, quantifying and understanding the flows of N between different sectors helps identify opportunities for increasing the efficiency of use of reactive N and reducing overall losses to the environment. In Denmark, much progress has been made in the absence of a national N budget, not least because the focus on − NO 3 leaching has demanded the use of a cross-sectoral approach. Strategies designed to achieve further reductions in losses of reactive N are likely to include increased recycling of N between economic sectors, increasing the need for the assessment of knock-on effects of interventions within the context of the national N cycle.
The development of national N budgets will be a useful method for disseminating knowledge of flows of reactive N outside the scientific community.